Skip to content

Start Here

Welcome to the Resilience Biomarkers Field Guide

This field guide documents the real journey of discovering and validating biomarkers for stress resilience in oyster aquaculture using RNA-seq meta-analysis. Rather than presenting an idealized workflow, this guide shares what actually worked, what didn't, and why—grounded in a year-long research effort analyzing multiple Crassostrea virginica (Eastern oyster) and Perkinsus marinus (parasite) RNA-seq datasets.

What This Guide Is

A Practical Chronicle

This is not a theoretical textbook. It's a field guide built from:

  • Real analysis decisions documented in 50+ notebook entries
  • Actual pivots and failures tracked across 25+ GitHub issues
  • Hard-won lessons about batch effects, leakage, and when methods break down
  • A working 6-gene classifier that emerged from systematic iteration

What You'll Find Here

Problem Framing

Understand what "resilience" means in this project, the datasets we're working with, and the biological and technical constraints that shaped our approach.

Pipelines

Two validated analysis paths:

  1. Stepwise Differential Abundance: Control vs. treated → resistant vs. sensitive
  2. Two-Step Classifier Pipeline: Reproducibility scoring → logistic regression

Both with explicit guidance on normalization, validation (LOSO), and common pitfalls (batch effects, leakage).

Year in Review

A chronological walk through the project from December 2024 to September 2025, highlighting:

  • Major methodological pivots (integrated → post-integrated analysis)
  • "Big Lessons" about data integration and trait definition
  • When to abandon an approach vs. when to iterate

Sources

Complete index of the GitHub issues and notebook posts referenced throughout this guide, organized by theme for easy exploration.

Glossary

Key terms and concepts defined in the context of this specific project.

How to Use This Guide

If you're new to biomarker discovery:
Start with Problem Framing to understand the research context, then explore the Timeline to see how approaches evolved.

If you're implementing a similar analysis:
Jump to Pipelines for validated workflows, then check Validation & Pitfalls for things that will bite you.

If you're troubleshooting an issue:
Use the Sources indices to find relevant discussions—chances are we hit the same problem.

If you're evaluating this approach:
Read the "Big Lessons" in the Timeline to understand when and why methods were abandoned or refined.

Core Finding: The 6-Gene Classifier

After extensive iteration, this project identified a 6-gene panel that successfully distinguishes tolerant from sensitive oyster phenotypes:

  • Emerged from post-data integration approach (combining datasets 1 & 5)
  • Validated using Leave-One-Study-Out (LOSO) cross-validation
  • Uses logistic regression to minimize feature set while maintaining accuracy
  • See Two-Step Classifier for full methodology

Attribution & Citation

This field guide is developed by Shelly A. Wanamaker as part of the Resilience Biomarkers for Aquaculture project.

If you use methods or insights from this guide, please cite:

Wanamaker, S.A. (2025). Resilience Biomarkers Field Guide. Resilience Biomarkers for Aquaculture Project. https://resilience-biomarkers-for-aquaculture.github.io/field-guide/

This guide is actively maintained and reflects work through September 2025. Check the Timeline for the most recent updates.